Hi John,
Statistically, 1 or 6 or 50 isn't enough to accurately characterise a
population exceeding a million.
Except.
I believe you will find the variation between examples, from
reputable producers, to be so small, that the performance can be
charaterised very well, within small limits.
But several examples are necessary. Single examples may vary a lot,
either way.
Another argument for trying new lenses before buying.
Tom
At 2000 June 6 - Tuesday 12:15, John A. Lind <olympus@xxxxxxxxxxxxxxx>
spoke about *Re: [OM] Photodo lens sharpness rat...* saying
> At 08:14 6/5/00 , Frank Ernens wrote:
> >Tom Trottier wrote
> >
> >> Perhaps we Oly users shouldn't be so sanguine about the quality of
> >> our lenses... [data from photodo]
> [snip]
> >Given that all the tests are conducted by a single person -
> >an ex-pro photographer, not an applied physicist - I would
> >question the methodology. Nowhere does it say what that was.
> [snip]
> >Testing one new sample of a lens doesn't tell you anything
> >about sample variation or whether the lens wears out quickly
> >and turns into a dog after a few months.
> [snip]
>
> Any lens test should have sufficient explanation about methodology to allow
> evaluation of method and replication of the test by others. This is basic
> to all scientific and engineering research, whether it is a "test" or an
> "experiment." Without it test results are confounded with unanswered
> questions about the test design.
>
> The existence of variation within a population has always bothered me about
> lens testing and published results. A sufficient sample size is required
> to capture the variation. I would want no less than a half dozen probably
> twice that many . . . and *not* one right after the other off the line . .
> . more if variation among the first sample is high. That is probably too
> much to ask. Lens tests should be viewed with a "grain of salt" and a dose
> of skepticism, the size of which should be based on revelation of test
> method and sample size. Judge that against the weight of user consensus
> about the particular lens. There are about a half-dozen criteria important
> about a lens; it's not all contrast, resolution and MTF.
>
> I applaud Gary Reese's efforts in his lens testing:
> 1. He explains how he tests the lenses thoroughly enough for a reasonable
> replication.
> 2. He makes an effort to test more than one example if he can get them,
> and gives a description of their condition.
> 3. He tests using different bodies and provides some analysis of
> differences noted between results for the same lens on different bodies.
>
> [take a bow Gary]
------------
Tom Trottier ICQ: 57647974 TomATrottier@xxxxxxxx
<<<<Abacurial Information Technology Consulting>>>>
400 Slater St. Suite 415, Ottawa ON Canada K1R 7S7
__o +1 613 291-1168 fax:594-5412 (877)247-8796
_ \< Vote for your favourite Olympus camera at
(+)/'(+) http://www.freevote.com/booth/fav_camera
< This message was delivered via the Olympus Mailing List >
< For questions, mailto:owner-olympus@xxxxxxxxxxxxxxx >
< Web Page: http://Zuiko.sls.bc.ca/swright/olympuslist.html >
|